A few openly available projects I carried out or contributed to
Multilingual Parallel Corpus Generator
Extract parallel sentences from Wikipedia articles existing in different languages. Part of my Master's thesis project.
Zure TTS
As part of the Codefest 2016, which took place in Donostia/San Sebastián in the Basque Country (4-8 July), a team of language specialists came together to create a new voice for the Aho TTS to let it speak with an Irish voice in the Donegal dialect of Irish (Gaelic). This was accomplished by collaboration between Euskal Herriko Unibertsitatea (University of the Basque Country), Trinity College Dublin, NUI Galway as well as Geraldine Uí Bhriain, Cathal Peelo and Ádám Varga.
Automated Essay Scoring
Automated essay scoring experiments for the Automated Essay Scoring Kaggle Competition.
Naïve Sound Classifier Tool
This tool can be used for naïve sound classification. It extracts MFCC coefficients from the sound files, and applies different state-of-the art supervised classification methods on them.
NLP Demos
Three simple online demos demonstrating basic NLP concepts and applications: Eliza, anagram search, and edit distance calculator/aligner
Metaphor Detection
This tool was used for experimenting with the automatic detection of metaphorical and/or idiomatic phrases in texts. The program constructs a full weighted graph for each sentence (the words being its nodes), where the weight is the Google Distance between two words, which serves as an automatic semantic distance measure. This version operated on an offline corpus, i. e. downloaded webpages.
